scientific article; zbMATH DE number 4029516

From MaRDI portal
Publication:3770503





In previous works the author introduced paraconsistent set theories based on first order paraconsistent logics, and proved that if the systems are nontrivial then the corresponding classical set theories are consistent [see \textit{A. I. Arruda} and the author, C. R. Acad. Sci, Paris 259, 2943- 2945 (1964; Zbl 0134.015), and the author, ibid. 260, 5427-5430 (1965; Zbl 0134.015)]. In a recent paper [Rev. Colomb. Mat. 19, 9-24 (1985; Zbl 0614.03052)] \textit{A. I. Arruda} has shown that the first formulations of da Costa's systems were trivial. The aim of this paper is to demonstrate that if the classical systems of set theory correlated with the authors's systems are consistent then the latter are nontrivial. After introducing the basic language of all systems of the paper (the primitive symbols are \(\supset\), \&, \(\vee\), \(\neg\), \(\forall\), \(=\), \(\in)\), the author defines the hierarchy \(NF_ i\), \(0\leq i\leq \omega\), of set \(theories.\) NF\({}_ 0\) is the system NF of Quine, \(NF_ 1\) is \(NF_ 0\) without the axiom \(\neg (A\&\neg A)\), the separation axiom subject to the restrictions that the formula F(x) must be stratified or is of the form \(x\not\in x\); among the postulates for the predicate calculus we have \(B^ 0\supset ((A\supset B)\supset ((A\supset \neg B)\supset \neg A\), \((A^ 0\&B^ 0)\supset ((A\supset B)^ 0\&(A\&B)^ 0\&(A\vee B)^ 0)\), \(\forall x(A(x))^ 0\supset (\forall xA(x))^ 0\) and \(\forall x(A(x))^ 0\supset ({\mathcal F}xA(x))^ 0\), with \(A^ 0\) being an abbreviation for \(\neg (A\&\neg A)\). The postulates are the ones of \(NF_ i\), replacing the symbols \({}^ 0\) in the above four axioms by the symbol \(^{(i)}\), where \(A^{(i)}\) constitutes an abbreviation for \(A^ 1\&A^ 2\&...\&A^ i\), with \(A^ i\) being \(A^{00...0}\) where \({}^ 0\) appears i times. Using the results which show that \(NF_ i\) \((1\leq i<\omega)\) contains \(NF_ 0\) and that for \(j<k\), \(NF_ j\) is stronger than \(NF_ k\) (for \(j,k=1,2,...)\), the author proves that if \(NF_ i\) \((1\leq i<\omega)\) is nontrivial then \(NF_ 0\) is consistent, and that \(NF_ i\) \((1\leq i<\omega)\) is inconsistent. Finally, he proves that if \(NF_ 0\) is consistent then \(NF_ 1\) is nontrivial and so, if \(NF_ 0\) is consistent then all inconsistent systems \(NF_ i\) \((1\leq i<\omega)\) are nontrivial. In this paper the author also shows that it is natural to introduce a hierarchy \(ZF_ i\), \(0\leq i\leq \omega\), starting with the ZF system; and proves that ZF is consistent if, and only if, \(ZF_ i\) \((1\leq i<\omega)\) is nontrivial. He ends the article making some remarks on the paraconsistent programme.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3770503)